CODEWITHSUNDEEP
python-3.xpandasnumpy
user032020
user032020

Reputation: 460

How to create variables based on multiple conditions?

In the below code, I need to create two variables, namely flag1 and flag2. They are created based on multiple conditions. I used np.select approach as below. However, I wonder what would be the other ways to do this? In my real work situation, there would be more conditions to create the flag. Any advices or suggestions would be great.

import numpy as np
import pandas as pd

start_date = '2020-04-01'
end_date = '2020-05-01'

d1 = {'customer type':['walk in', 'online app', 'phone app', 'referral'], 'office visit': 
['location1','location1','location1','location1'],'date1':['2020-04-17','2020-05-17','2020-03- 
01','2020-05-01'],'date2':['2020-05-18','2020-04-18','2020-04-03','2020-05-19']}

df1=pd.DataFrame(data=d1)

con1 = [ (df1['date1'] >= start_date ) & (df1['date1'] < end_date )]
result1 = ['yes']
df1['flag1'] = np.select(con1, result1)

con2 = [ (df1['date2'] >= start_date ) & (df1['date2'] < end_date )]
result2 = ['yes']
df1['flag2'] = np.select(con2, result2)

Upvotes: 0

Views: 485

Answers (1)

CobraPi
CobraPi

Reputation: 392

You could use a dictionary and dynamically update the keys to the variable names and add the corresponding value of the variables.

For example:

import numpy as np
import pandas as pd

start_date = '2020-04-01'
end_date = '2020-05-01'

flags = dict()


flag_string = 'flag'

# This creates the strings flag1 and flag2 automatically
for i in range(1, 3):
    # concatenate the flag_string with the index of the loop
    flags[flag_string + str(i)] = flag_string + str(i)

print(flags)

d1 = {'customer type': ['walk in', 'online app', 'phone app', 'referral'], 
      'office visit':  ['location1','location1','location1','location1'],'date1':['2020-04-17','2020-05-17','2020-03- \
01','2020-05-01'],'date2':['2020-05-18','2020-04-18','2020-04-03','2020-05-19']}

df1=pd.DataFrame(data=d1)

con1 = [ (df1['date1'] >= start_date ) & (df1['date1'] < end_date )]
result1 = ['yes']
df1[flags['flag1']] = np.select(con1, result1)

con2 = [ (df1['date2'] >= start_date ) & (df1['date2'] < end_date )]
result2 = ['yes']
df1[flags['flag2']] = np.select(con2, result2)

This is how you can substitute dictionary values as variables. I've also included a for loop that builds your flag dictionary.

Upvotes: 1

Related Questions